Scribe¶
Why Scribe¶
Modern systems do not fail in one place. A single user action can cross services, queues, retries, feature flags, and third-party calls. Traditional logging turns that into dozens of partial lines, each easy to write and hard to query.
Scribe is built around the argument from loggingsucks.com: useful logs are not just structured, they are context-rich. Instead of scattering breadcrumbs everywhere, Scribe gives you primitives for both shapes you actually need:
Notefor one-off events that stand on their own (Old way, if you still insist)Scrollfor building a wide event over the lifetime of an operation, then sealing it as aSealedScroll
That pushes logging toward "what happened to this request or workflow?" instead of "what line of code ran next?". The result is fewer events, better context, and logs that are easier to filter, correlate, and analyze.
Read Next¶
- Getting Started for setup and first usage
- API Concepts for terminology and data model
- Lifecycle and Delivery for buffering, margins, and shutdown